FlexQueries

Features

• AI brand visibility lookup. Look up a brand or domain across supported ChatGPT and Google AI Overview mention data, or run a Google AI Mode query, then inspect the prompts, cited pages and source domains behind each mention. Official source 1 ↗
• Exact-question AI answer tracking. Save up to 25 exact questions and five competitors, collect answers from ChatGPT, Gemini, Google AI Overview and AI Mode (Claude and Perplexity as labelled API answers) with one to three samples, and track written mentions, citation coverage and share of voice per engine, manually or on a daily, weekly or monthly schedule. Official source 1 ↗ · Official source 2 ↗
• Model interpretation of AI answers. An optional, separately priced model call labels per-brand sentiment and numbered recommendations in a saved answer, with evidence spans that must match the saved text; one interpretation has a $0.05 ceiling. Official source 1 ↗
• Flex AI Agent. A dashboard chat agent answers research questions with project-bound tools and connected reports, and exports answers as PDF, Word or CSV tables. Its model steps are metered, with a default $10 maximum per request adjustable from $0.05 to $50. Official source 1 ↗ · Official source 2 ↗
• Content Studio briefs and draft scoring. Content Studio creates research-backed content briefs, scores drafts, applies keyword clusters and suggests internal links; the REST API also generates paid SERP-backed content briefs. Official source 1 ↗ · Official source 2 ↗
• Keyword research and SERP history. Expand a seed topic with volume, difficulty, CPC, intent, autocomplete and trends, inspect current results and up to twelve months of historical SERP snapshots, and save tagged shortlists. Official source 1 ↗
• Competitor, backlink and marketplace research. Research domain overviews, ranking keywords, competitor gaps and backlink profiles, plus Apple App Store, Google Play and Amazon keywords and competitors. Official source 1 ↗
• Rank tracking and site audits. Track keyword positions by market and device on manual, weekly or daily schedules, and run site audits of up to 500 pages with 20 Lighthouse checks. Official source 1 ↗ · Official source 2 ↗
• Hosted SEO MCP server. One OAuth connection gives an AI client project-aware tools for research, audits, rank tracking, AI visibility, Content Studio, reports and Search Console, with per-connection project, write and action/daily/monthly/lifetime spending limits. Official source 1 ↗
• Page planning and opportunities. Keep an inventory of up to 1,000 known URLs and unpublished page ideas with target keywords, and review audit, Search Console and GA4 evidence per page; organizing plans does not use credits. Official source 1 ↗

Pricing & free limits

Billing details

Example usage prices: keyword report (225 rows) $0.15, backlink overview $0.20, 500-page site audit $1.30, local Maps lookup $0.05, two-platform AI visibility run $0.85, Prompt Explorer response up to $0.90, 50 weekly tracked keywords $2.18 a month. Prices vary with scope. Official source 1 ↗

Exact-question AI tracking: 50 short ChatGPT/Gemini checks cost up to $0.50 with queued Standard delivery or $1.00 Live; an AI brand check across ChatGPT and Google is $3.05. Official source 1 ↗ · Official source 2 ↗

Each paid action reserves a temporary hold that can exceed the estimate, then settles actual usage; provider-backed actions round up to the next $0.05. When the balance runs out, paid actions stop rather than overdraw. Official source 1 ↗ · Official source 2 ↗

Unused purchased balance can be refunded on request; amounts already consumed are non-refundable except for confirmed billing errors. Credits cannot be transferred or cashed out. Official source 1 ↗

Fees charged by your own AI client or model provider are separate; Flex AI Agent's own model work is metered from the FlexQueries balance. Official source 1 ↗

Free access & limits

Eligible verified new accounts get $1.00 of promotional credit once, with no card; after that, all paid research needs purchased credits. Official source 1 ↗

The public free SEO tools are not unlimited: they require sign-in and draw on the same balance. Official source 1 ↗

Saved-data reads, exports, estimates, Search Console and GA4 reports, and eligible cache hits do not use credits. Official source 1 ↗

Best for & limitations

Check whether ChatGPT, Gemini or Google AI answers mention and cite your brand for a set of buying questions, and follow the results over time. Official source 1 ↗

Ask Claude or another MCP client to research keywords around a topic, compare demand and difficulty, and save a tagged shortlist in your project. Official source 1 ↗

Turn a site audit into a prioritized fix list with an AI agent instead of pasting exported rows into a chat. Official source 1 ↗

Pull keyword, competitor and saved project data into an internal app or agency reporting workflow through the REST API. Official source 1 ↗

• AI visibility covers sampled, supported data only. Reports cover the supported surfaces and the questions you save, not every AI conversation or platform; a missing result does not prove a brand never appears. Official source 1 ↗ · Official source 2 ↗
• Claude and Perplexity results are API answers, not app results. Only ChatGPT, Gemini, Google AI Overview and AI Mode are collected from search interfaces; Claude and Perplexity entries are model/API answers. Official source 1 ↗
• Sentiment and recommendation labels can be wrong. Optional interpretation produces model-generated labels, not fact checks, and is a controlled-rollout module like cited-page research. Official source 1 ↗ · Official source 2 ↗
• Connected agents can spend without dashboard confirmation. Paid MCP calls draw on the shared balance without a dashboard prompt, so per-connection limits and client tool confirmations matter. Official source 1 ↗
• Empty results can still be charged. A successful lookup that finds no matches is billable, and successful checks in a partly failed run keep their charge. Official source 1 ↗ · Official source 2 ↗
• No publishing to your site. FlexQueries plans pages and suggests work but does not publish content; you publish through your own CMS. Official source 1 ↗